Abstract: In some wind farms, the wind power generator may aim the wind direction no so accurately, and the wind power generators located in leeward of other wind generators are impacted by air flow of other wind power generators and can not aim the wind direction exactly. Aiming at above-mentioned problems, the KHC algorithm, a new control algorithm, in which the optimized Karman filtering is combined with hill climbing method, was put forward. The core contents of the algorithm are that using the power detection to aim the wind direction and contral high accurately， and using the Kalman filtering to predict the wind vector to avoid the frequent and fatigue action of the yaw mechanism and the interferences. By simulation analysis, the feasibility of the algorithm was verified; and the wind power generation efficiency had been increased greatly.

Key words: wind power generation unit, yaw control, kalman hill climbing control (KHC) algorithm, power detection, prediction of wind vector